The following meanings for the abbreviations used in this specification apply:    AAA: AA (Authentication and/or Authorization) Answer    AAR: AA (Authentication and/or Authorization) Request    APN: Access Point Name    AVP: Attribute Value Pair    CCA: Credit Control Acknowledgement    CCR: Credit Control Request    IP: Internet Protocol    IP-CAN: Internet Protocol Connectivity Access Network    NAT: Network Address Translation    PCEF: Policy and Charging Enforcement function    PCRF: Policy and Charging Rule Function    PDN-GW: Packet Data Network Gateway    P-GW: Packet Gateway    QoS: Quality of Service    QCI: QoS Class Identifier    RAA: Re-Auth-Answer (RA-Answer)    RAR: Re-Auth-Request (RA-Request)    UE: User Equipment
Embodiments of the present invention relate to providing additional services to a single IP stack UE. A “Single IP stack” UE is defined as a terminal, which supports only one IP-Address at a given time. Most of the terminals support single IP stack. Moreover, a number of operators use single APN concept, when all connectivity to a UE is provided via a single APN and NAT is used to access different address realms, where application servers are deployed. Thus, in some deployments UEs will get only a single IPv4 address even if they can support multiple IP connectivity.
3GPP in TR 23.975 (in more detail, TR 23.975v11.0.0: “IPv6 Migration Guidelines”, 3GPP, (06-2011)) has defined a network deployment, where the services are defined in one address realm and the UE and core entities are in another address realm, connected via a NAT. This deployment allows a network provider to have services in a different address realm which can be managed independently. FIG. 2 shows one of such deployment taken from TR 23.975 (section—Annex A):
In particular, FIG. 2 shows a situation in which a UE is located in an address realm 1 and uses an address AP1 in this address realm. That is, the UE and a core network is located in address realm 1. The UE is connected to a PDN gateway. The PDN gateway is connected (via interface SGi) to a NAT, which provides a connection to an address realm 2. The NAT performs an address translation between the two address realms, i.e., for the UE, which uses AP1 in address realm 1, an address AP2 is assigned in address realm AP2. By means of this address, operator's IP services can be reached, which include an AF. Moreover, a PCRF controls policy decisions etc. in connection with the services provided for the UE in address realm 2. The PCRF is connected to the operator's IP services (AF) via interface Rx, and is connected to the PDN gateway in address realm 1 via interface Gx.
This situation can lead to problems, as will be described in the following:
The single stack UE is allocated an IP Address (AP1) during initial attach, by the core network in address realm 1. In multi-realm deployments, when such a user tries to access the services which is situated in the address realm 2, this can create a problem, as the IP Address of this UE is translated from AP1 to AP2 by the NAT which separates the address realm 1 from the address realm 2. The important issue is where a user is on single APN. Here, the user is already accessing some services with his IP-Address (AP1). When the user intends to access the services of address realm2, an IP-Address (AP2) is to be used. So, such single stack UE cannot access both the services as the device only supports one IP-Address for a given IP-CAN Session. Also, the IP-address provided on Rx interface (AP2) will never match with the IP-Address provided on Gx (AP1). The only available solution until now is that the IP-Address (AP2) is assigned to the IP-CAN session again, which may result in terminating and re-establishing the existing services on such UE.
This situation is illustrated in FIG. 3. As mentioned above, the single stack UE is assigned an IP-Address (AP1) during initial attach. The UE has created a dedicated bearer (QCI=6) for browsing internet services. Now the UE wants to access operator services simultaneously, which is in address realm 2, that requires address AP2. Now, if the UE is assigned an address AP2, then this would result in disruption of service on QCI=6, as the UE is single stack which allows him to be on one address always.